Research On Rail Crack Detection Method Based On Ultrasonic Guided Waves

Rail fatigue crack and other internal damage can induce rail break accident,which is one of the key points of track maintenance.At present,the rail flaw detection car based on ultrasonic body wave is widely used for crack detection at home and abroad.However,due to the large attenuation of ultrasonic body wave signal,there is a detection blind area at the rail base of the flaw detection vehicle.Therefore,this dissertation proposes a rail flaw detection method based on ultrasonic guided wave,which can detect the whole cross-section characteristics of rail in a long distance,and realize the full section detection of rail crack damage.Based on the in-depth study on the propagation characteristics of ultrasonic guided waves in rail,the key issues of rail crack detection,such as mode selection,excitation and crack location,are studied in this dissertation,which lays a theoretical foundation for the detection of rail full section cracks.Firstly,aiming at the problem of selecting the detection mode of the crack in different regions of the rail,a method of selecting the best detection mode based on the crack sensitivity of guided wave mode is proposed.Based on the analysis of the dispersion characteristics and mode characteristics of guided wave modes,the crack area energy and crack reflection intensity indexes are established,and the action mechanism of different cracks on modes is studied.Through the analysis of the dispersion characteristics of guided wave mode group velocities in rail,the nondispersion characteristic index of mode is further introduced,and the modes which affect the detection accuracy due to serious dispersion are eliminated.The simulation and experimental results show that the guided wave mode selected by this method can accurately identify the rail crack,and it lays a foundation for the research on the accurate location method of rail crack.Aiming at the problem of target mode excitation,a guided wave mode excitation method based on graphical modal analysis method is proposed.The complex analytical problem is transformed into image processing problem by transforming the modal information of guided wave into RGB image.Then K-means clustering algorithm is used to classify the high-frequency complex modes of rail.Combined with histogram method,edge extraction algorithm and global threshold binarization image processing methods,the optimal mode excitation mode of rail crack detection in different regions is determined.The simulation results show that the method can excite the guided wave signal with the target mode as the main component in the rail,and solve the problem that the mode excitation mode cannot be determined intuitively and quickly due to the multi-modal characteristics of the guided wave and the complexity of the vibration mode.It provides a technical means for the detection and location of rail cracks.Aiming at the problem of long-distance detection and location of rail cracks,a rail location algorithm based on the single mode extraction algorithm is proposed.Firstly,all guided wave modes and their relative energy amplitudes in the echo signal are identified by using the three-dimensional array guided wave signal analysis method.On this basis,the reflected echo time domain signals of each guided wave mode are separated by using the mode tracking method,and the actual arrival time of each guided wave mode is obtained through time-frequency analysis.Finally,combined with the real-time calibration method of group velocity of elastic modulus matching,the detection and location of rail crack is realized.The simulation analysis and the experimental results show that the single mode extraction algorithm solves the problems that the frequency aliasing is difficult to distinguish the different mode echo,and the crack defect cannot be subdivided and positioned in the detection range,which improves the accuracy of crack detection and location.The research results show that the method proposed in this dissertation can effectively realize the selection of rail full section crack detection mode,the excitation of guided wave mode and the detection and location of crack,which provides a new idea for realizing long-distance rail crack detection without blind area.